Abstract
In this paper, a new method by the regression tree model for the quantitative habitat evaluation is proposed. We compared the regression tree model with Suitability Index (SI) models of the Habitat Evaluation Procedure (HEP) or that of the Physical Habitat Simulation Model (PHABSIM). For our comparison, we used habitat characteristics of fish spawning beds in floodplain pools. Our results indicated that the error of regression tree model was smaller than that of SI models, and suggested that the regression tree method can be used for the habitat suitability calculation method of the HEP or the PHABSIM.
